In the US alone, it is estimated that lung cancer caused a total of 161,900 deaths in 2000. The 5-year survival rate in the US is 13% when all stages are considered. Substantial evidence suggests that early detection of lung cancer may reduce mortality for patients with stage T1 NxMx. High-resolution CT can detect lung nodules when they are still small. But it is currently difficult for radiologists to distinguish those that are malignant from those that are benign. We propose to develop a computer-aided diagnosis (CAD) system to assist radiologists in the diagnosis of lung cancer in thoracic computed tomography (CT). The CT-CAD system will enable radiologists to robustly measure and analyze the suspicious lesions, effectively visualize the region of interest, and improve differentiation between malignant and benign nodules. Specifically, this SBIR phase II effort includes (1) continuous development of dedicated lung cancer diagnostic tools, (2) development of a complete CT-CAD system for accurate analysis of small lung lesions, (3) evaluation of the proposed system functions and integration of them in a CAD-specific 3D visualization platform; and (4) evaluation of the radiologists performance with and without using the CT-CAD system using the MRMC-ROC study protocol. The successful development of the proposed CT-CAD system can significantly benefit the existing CT lung cancer diagnosis in two-fold: (a) improving sensitivity for lung cancer detection and (b) improving the accuracy of lung cancer diagnosis in CT and leading to reduction of unnecessary biopsies. We expect that this CT-CAD can facilitate remote reading by experts and will also increase radiologists' efficiency in reading large image arrays and reducing the observer variations in thoracic CT image interpretation.